Nero's 'Architects', Severus and Celer, and Residence Patterns in Rome

N ero’s ‘Architects’, Severus and Celer, and Residence Patterns in Rome* C hrister Bruun 1. T w o com panions found in a neglected piece o f evidence N e ro ’s fam ous G olden H ouse was planned by tw o m a g is tr i e t m a c h in a to r e s called Severus and C eler (Tac. A n n . 15 .4 2 ).1 The im perial palace they designed continues to arouse interest, but they them selves have received little attention. T his is undoubtedly due to the fact that, except for T acitu s’ b rie f m ention, scholars previously had no o ther inform ation on the tw o ‘architects and engineers’, as m o d em translations conventionally render their pro fessio n .2 T here is how ever a hitherto n eglected source that provides som e new inform ation, w hile at the sam e tim e raising som e new questions. T his source is a stam p on a lead pipe ( fis tu la ) from Rom e w hich H einrich D ressel re a d thus (see C IL X V 7393): ΑΝΤΟΝΙΑΕ I[.]VI C[...]ET DVVM [,.]LIORVM SEVERI ET CELERIS IIII My sincere thanks to Dr. Giorgio Filippi of the Vatican Museum for generous assistance in connection with my work on the lead pipe stamps in the museum; this study is part o f that larger project. I am much indebted to Dr. Mikael Lindström, my old friend and onetime fel­ low student in the Department of Mathematics o f Abo Akademi, for his expertise. The paper also benefitted from the comments o f the two anonymous referees o f SC I. Finally, I am grateful to Fred Unwalla for improving my English. All remaining errors are my own. Research on this paper was facilitated by a Standard Research Grant from the Social Sciences and Humanities Research Council o f Canada, which is most gratefully acknowledged. The OCT text o f Tac. A n n. 15.42 reads: C eterum N ero usu s est p a tr ia e m in is extru xitq u e do m u m in q u a h a u d p ro in d e g em m a e et au rum m ira culo essent, so lita p rid e m et luxu v u l­ gata, q u a m a rva et sta g n a et in m odu m so litu d in u m h in c silva e inde a p erta sp a tia et p rospectu s, m a g istris et m a ch in a to rib u s S evero et Celere, q u ib u s ing en iu m e t a u d a cia era t etia m q ua e n a tu ra d en eg a visset p e r artem tem ptare et virib u s p r in c ip is inludere. na m q ue ab lacu A ve rn o n a viga b ilem fo s s a m usq ue a d o stia Tiberina d ep ressu ro s p r o m is e r a n t ... ‘How­ ever, Nero turned to account the ruins of his fatherland by building a palace, the marvels o f which were to consist not so much in gems and gold, materials long familiar and vulgarized by luxury, as in fields and lakes and the air of solitude given by wooded ground alternating with clear tracts and open landscapes. The architects and engineers were Severus and Celer, who had the ingenuity and the courage to try the force o f art even against the veto o f nature and to fritter away the resources o f a Caesar. They had undertaken to sink a navigable canal running from Lake Avernus to the mouths of the Tiber’ (transi. J. Jackson, LCL). No information on Severus and Celer is found in, e.g., Η. Fumeaux, The A n n a ls o f T acitu s2, (ed. w. introduction and notes), Oxford, 1907, 370; Ε. Koestermann, C orn eliu s Tacitus A n ­ n alen IV (Buch 14-16), Heidelberg, 1968, 246. Whether ‘achitects and engineers’ is a proper translation o f m a g istri et m a ch in atores is not o f major relevance for the argument o f this paper; see however below, section three, for the debate on the question. S cr ip ta C la ssica Isra elica vol. XXVI 2007 pp. 73-86 74 NERO’S ‘ARCHITECTS’ and ‘in parte aversa’ on the fistula'. C. IVLIVS PINYTVS FEC. The suggestion that w e are dealing here with N e r o ’s ‘architects’ has to m y k n ow led ge never been advanced b efore.3 D ressei in C IL X V a d loc. restricted his com m ents to pro­ posing, with som e doubts, that w e m ight be dealing with a w om an called A n to n ia L iv i.4 H e also suggested the reading [ lu ] Ι io ru m or [ A e jlio r u m for the g e n tilic iu m o f Severus and C eler.5 2. T he statistical p robability underlying the identification M y grounds for suggestin g that w e are dealin g here w ith N e r o ’s ‘architects’ are statisti­ cal. W e need to consid er the probability that — in any random pairing o f tw o c o g n o m in a from R om e — one should end up with precisely ‘Severu s’ and ‘C eler’. A s it turns out, the chances are extrem ely rem ote. B y far the m ost plausib le explanation for the pairing in C IL X V 7393 o f the tw o sam e nam es, Severus and Celer, that T acitus jo in s together, is that w e are not, in fact, dealing with a coincid ence: th ese m en m ust be the very sam e ones w ho are m entioned in Tac. A n n . 15.42. Som e num bers and som e sim ple m athem atical calculations w ill bear this out. In the inscriptions published in C IL VI, w hich surely are representative en ou gh o f the epigraphical sources su rvivin g from the ancient city o f R om e, and w h ich I use here for the sake o f con ven ien ce, there are, according to a recent estim ate, 5 4 ,0 0 0 in div id uals.6 N o t See most recently on Celer (Tac. Ann. 15.42) in N P 2 (1997), 1049 Nr. 4 (W. Eck) ‘mög­ licherweise ist er mit dem a rch itectu s Celer in P. Ryl. 608 = C P L 248 identisch (1.28f.)’; W. Eck in Lexico n T op o grap h icu m U rbis R o m a e (L T U R ) II, Roma, 1995, 78 s.v. ‘domus: []lius Celer’ (with no connection to the passage by Tacitus). Nothing o f relevance for the identification proposed here is found in L. Guerrini, O d e r ’, EAA II (1959), 456; P IR 2 C 619; P IR S 452; R E ΙΙ.2Α (1923), 1937 s.v. ‘Severus Nr. Γ (Stein); R E III.2 (1899), 1870 s.v. ‘Celer Nr. 14’ (Fabricius). None of the editions o f book 15 o f Tacitus’ a n n a les that I have consulted make this connection; cf. in particular n. 2 above, and also, e.g., Α. A rid (ed.), Tacito, A n n a li, Torino, 1952, 558; F. Römer (ed.), in W S Beiheft 6, Wien, 1976; Ρ. Wuilleumier (ed.), in Tacitus, A n n a le s iv (Collection G. Budé), Paris, 1978; R. Oniga (ed.), Tacito II. A n n a li , Torino, 2003, 1526-7. The same doubts were expressed by M.-Th. Raepsaet-Charlier, P ro so p o g ra p h ie d es fe m m e s de l'o r d re sé n a to ria l (I-II s.), Lovanii, 1987, 91 no. 75, but not by W. Eck, ‘Die fis tu la e a qu ariae der Stadt Rom: zum Einfluß des sozialen Status auf administratives Handeln’, in idem, D ie V erw a ltun g d es R ö m isch en R eiches in d er H o hen K a iserzeit. A u sg ew ä h lte u n d erw eiterte B e iträ g e 2, Basel, 1998, 245-77, esp. 261. It is worth noting in relation to the name o f the p lu m b a riu s C. Iulius Pinytus that one Pinytus Divi Aug(usti) l(ibertus) appears in A E 1965, 335; after his manumission, he would naturally have been a ‘C. Iulius’. There might be a relationship o f some kind between him and the p lu m b a r iu s. No significant meaning can be attributed to the figure 1111 (‘four’) which appears in the stamp; see on numbers on Roman fis tu la e in general C. Bruun, The W ater S u p p ly o f Im p eria l Rom e: A S tu d y o f R o m a n Im p eria l A d m in istra tio n , Helsinki, 1991, 44-51. I take this figure from Ο. Salomies, ‘People in Ostia. Some Onomastic Observations and Comparisons with Rome’, in C. Bruun and A . Gallina Zevi (eds.), O stia e P o rtu s n elle loro rela zio n i con R o m a (Acta IRF 27), Roma, 2002, 135-159, esp. 136. CHRISTER BRUUN 75 all o f these individuals have a cogn om en , and so I w ill use the round number 5 0 ,0 0 0 (individuals with a cogn om en ) for m y calculations. The indices to C IL VI contain references to all the c o g n o m in a in that corpus. M y c a l­ culations sh ow that, in C IL VI ‘Severu s’ appears som e 3 5 0 tim es, and ‘C eler’ som e 145 tim es.7 T h ese tw o nam es are certainly not rare,8 but it is easy to see that o n e nam e c h o ­ sen at random from am on g the nam es recorded in ancient in scriptions from R om e has only a probability o f 3 5 0 /5 0 ,0 0 0 = 0.7% o f bein g ‘S everu s’, w h ile the chance that w e m ight ch o o se som eon e called ‘C eler’ is less than h a lf o f that, for 1 4 5 /5 0 ,0 0 0 = 0.29% . The probability that any tw o ancient Rom an nam es that w e happen to ch o o se w o u ld be precisely Severus and C eler are, how ever, incom parably sm aller. T he probability that tw o particular nam es out o f a total o f 5 0 ,0 0 0 nam es m ay be random ly com bin ed is c a l­ culated according to the form ula for M ultinom inal Distribution:9 Total names = 50,000 The probability that the first name chosen at random is ‘Severus’ = 350/50,000 The probability that the second name chosen at random is ‘Celer’ = 145/50,000 The probability o f ending up with one Severus and one Celer when combining two names at random from the pool o f Roman names: Ρ (Severus, Celer) = 2! χ (350 / 50,000] χ £145 / 50,000] = 2 χ 50.750 « Ι Ι! Ι! (50,000x 50,000) 25,000 In fact one m ight argue that the chances are on ly h a lf as great, for T a citu s’ expressio n is S e v e ro e t C e le re , and I w ould su ggest that it is hig hly sign ifican t that w e have the sam e order in the inscription. W hen p eop le b ecom e fam ous and are talked o f as a pair — and I argue that this is what happened to N e r o ’s ‘architects’ — very often a standard order o f reference is established: Laurel and Hardy, G ilbert and Su llivan, R odgers and H am m erstein, Frank and Jesse Jam es, W ilbur and O rville W right, Lennon and M cCartney etc., to give som e exam ples from m od em culture.10 There m ay be particular reasons w hy so m eone w ould w ish to change the order and refer to, for instance, M cC artney and 7 8 9 10 See L. Vidman, C IL VI 6.2. In d ex co gn om in u m , Berolini, 1980. These figures are, in proportion, not too far from the frequencies presented in I. Kajanto, The La tin C o gn om in a, Helsinki, 1965, 248 and 257. Based on his much wider survey of epigraphical material from all over the Roman world, Kajanto gathered a total o f some 133,000 Latin co g n o m in a (27), o f which some 121,000 belonged to the imperial period (29), the period covered by CJL VI. Among the names from the imperial period, Severus ap­ peared 1,214 times, and Celer 516 times (including senators). This gives a proportion, out of Kajanto’s onomastic material, o f 1.0% for Severus and 0.43% for Celer. In Rome, the fig­ ures are 0.7% and 0.29%, respectively. It is worth noting that Kajanto collected only Latin co g n o m in a, while the Indices o f C IL VI include Greek co g n o m in a too, which are even more numerous than the Latin ones. Therefore it would seem that among Latin co g n o m in a, Severus and Celer are somewhat more frequent in Rome than in the Empire overall; Severus is about twice as common in both cases. See ΜΉ. Spiegel, J. Schiller and R.A. Srinivasan, P ro b a b ility a n d S ta tistics 2 (Schaum’s Outlines), New York, 2000, 118, 138-9. Compare an example from Rome: we find the same order between two brothers in C IL XV 7760 O rfiti et P iso nis, and in C IL VI 9830 = IL S 7388 O rfiti e t P iso n is lib. et p ro c. 76 NERO’S ‘ARCHITECTS’ L en n o n ," but generally, I think that m ost readers who are fam iliar w ith these nam es w ould agree that it feels odd to say ‘H am m erstein and R odgers w rote T h e S o u n d o f M u s ic ’ or ‘H ardy and Laurel w ere great com edians’. A rguably we are then looking for ‘Severus and C e le r’, in that order, and can expect this com bination in 0.002% o f the cases w here we encounter only tw o nam es, o r about once every 50,000 pairs o f nam es from R o m e.1112 In case any doubts still linger, regardless o f the statistical calculations presented above, we shall cast a b rie f glance at the em pirical m aterial at our disposal. N eedless to say, the pair Severus and C eler appears now here else in the epigraphical m aterial that I have encountered. It m ust, how ever, be said that there are a num ber o f inscriptions from Rom e in w hich the tw o nam es Severus and C eler are both present. T hese cases are all inscriptions consisting o f long lists o f m em bers o f various organizations in R om e, m ostly m ilitary u n its.13 It is obvious that the probabilities for the occurrence o f ‘S everus’ and ‘C eler’ change when we are dealing with a group o f som e 680 nam es (as in C IL V I 200) or 350 nam es (as in C IL VI 975). In such large sam ples it w ould be m ore surprising if the nam es w ere not present, especially as both nam es clearly w ere favoured by m ilitary m e n .14 It is also im portant to note that the fact that tw o nam es appear scattered som ew here in these ex­ tensive lists o f citizens o r soldiers im plies nothing about the relation o f the bearers o f these nam es, except that they w ere fellow m em bers o f an organization o r c o m m ilito n e s in a very general sense. T he m ilitary lists are norm ally divided in colum ns, w hich are structured according to a system w hich is not alw ays alphabetical and m ust have som e­ thing to do w ith service conditions. O nly if the nam es Severus and C eler appeared in 11 12 13 14 Cf. the attempt by Paul McCartney to change the order o f the names on the copyright for some Beatles songs, from the standard ‘Lennon — McCartney’ to ‘McCartney — Lennon’, as reported in the international press in the spring of 2003. The probability in this case is calculated as follows: (a) the chance o f drawing Severus as first name is 350/50,000; (b) the chance o f drawing Celer as second name is 145/50,000 (the denominator in the second case should really be 49,999 but since we are operating with an estimate and the difference is minimal in any case, I use 50,000 for the sake o f simplicity); (c) the combined probability of drawing first Severus and then Celer is (350/50,000) χ (145/50,000) = 50,750 / (50,0000 χ 50,000). This gives a 0.00203% probability o f encoun­ tering our pair, or (roughly) 1:50,000. Such inscriptions are: C IL VI 200 (the trib u s S uccu sa n a [co rp u s] iu n io r[u m ], Severus once, Celer twice); VI 975 (an inscription of the vicom ag istri, Severus once, Celer once); VI 1056 (coh. Ι vig ilu m , Severus nine times, Celer three times); VI 1057 (c o h . V vig ilu m , Severus 11 times, Celer 4 times); VI 1058 (coh. V vigilu m , Severus eleven times, Celer once); VI 1063 (vigiles, once Severus, once Celer); VI 2071 (A cta o f the Arval Brethren, Severus 8 times, Celer once; the Index o f CIL VI does not give the whole picture here. As one can see in J. Scheid, C o m m en ta rii fr a tr u m A rva liu m q u i sup ersun t, Rome, 1998, 140-1 no. 53, in this fragment from 84 CE, two senators, Ti. Tutinius Severus and L. Pompeius Vopiscus Arrun­ tius Catellius Celer, are officiating, and each is mentioned several times); VI 32515 (p ra eto ria n i and u rb an ician i, Severus 4 times. Celer once; the name Severus also appears nine times in the inscription as part of the consular dating S ev ero II [consule])·, VI 32520 (p ra e to ria n i , Severus nine times, Celer once). L.R. Dean, A S tu d y o f the C o g n om in a o f S o ldiers in the R o m a n L e g io n s , Princeton, NJ, 1916, 19-20, 51-2, 61-2. CHRISTER BRUUN 77 very c lo se proxim ity in such a colum n w ould one be entitled to con clu d e that a relation m ig h t have existed betw een the tw o m en. Such cases are e x ceed in g ly rare.15 A ll in all, it is clear that the pairing o f Severus and C eler in Tac. A n n . 15.42 and in C IL X V 7393 is an excep tion al and statistically significant occurrence that ough t to be explained. W e n eed to look for a connection jo in in g these tw o nam es. But the nam es have no etym ological or sem antic connection that m ight explain the pairing. N o r do the nam es Severus and C eler have any connection in m yth, legend or history, as far as on e can s e e .16 Thus I can think o f no other explanation w hy the nam es Severus and Celer are found paired tw ice other than that w e are dealin g with the sam e individ uals in both c a se s.17 3. C eler as architectus in Egypt? B efore discu ssin g the im plications o f this n ew source on Severus and Celer, another piece o f ev id en ce m ust be taken into account. A n Egyptian papyrus written in P anopolis during the secon d part o f the first century C E ,18 that is, during the period w hen N e r o ’s ‘architects’ w ere active, w as discussed in detail som e tw enty years ago by Hannah C o t­ ton. The papyrus contains a letter o f recom m endation to an im perial procurator Ti. Claudius H erm eros, written by on e ‘[-]ius C eler’, w ho signs the letter C e le re a rc h ite c to . Cotton suggested that w e are dealing here with one o f the ‘architects’ o f N e r o ’s G old en 15 16 17 18 In CIL VI 1056, in the cen turia lu ve n is we find one Octavius Celer as no. 104 and one Spu­ rius Severus as no. 105. In the same cen turia we also have no. 35 Marius Severus and no. 78 Calpurnius Celer. Otherwise, only twice do we find both names within ten places: in C IL VI 1058, in the cen tu ria A n tu lli as no. 95 Τ. Valerius Celer, as no. 105 C. Albicius Severus; and in C IL VI 32520, Τ. Lartius Severus from Tuder in col. VI.b.4 (enrolled in 143 CE), and L. Caninius Celer from Luca in col. VI.b.9 (enrolled in 144 CE). Η. Solin, N a m en pa are. E in e S tu d ie z u r rö m isch en N a m en g e b u n g , Helsinki, 1990 and Η. Solin, 'Coppie di nomi’, in Μ. Pani (cd.), E p ig ra ß a e territorio. P o litica e società, terni d i a n tich ità rom a n e 4, Bari, 1996, 353-69, has shown that names which belong together in a mythological, religious or historical context, like Amphio and Zethus, sometimes appear in the same family or slave fa m ilia . Other possible reasons for the pairing o f certain names with more than average frequency, such as end-rhyme ( Vo lum ni d u o Verus et S everus in C IL XIV 2495a), or consonance {MM. C occei V erecundus et Verus in C IL VIII 12665 = IL S 1550) are not present in the case of Severus and Celer. Hypothetically, it could be argued that precisely the pair mentioned by Tacitus served as inspiration for a father or a slaveowner with particular ambitions for two young boys. It is known that in some professions in Rome, notably among physicians and actors, there were ‘trade names’ that probably were taken up by professionals at some stage o f their careers; see Η. Solin, ‘Die sogenannten Berufsnamen antiker Ärzte’, Ρ. v. d. Eijk et al. (eds.), A n ­ cient M ed icine in its S o cio -C u ltu ra l C o n text I, Amsterdam, 1995, 119-42; idem, ‘Nochmals zu Berufsnamen bei antiken Ärzten’, A c ta C la ssU n ivS cien tD eb recen sis 34-5 (1998-9), 38993. I attribute no importance to Dressel’s dating o f the fis tu la to the late first or early second century CE ( C IL XV, a d loc.). The precise dating of Latin inscriptions on paleographical grounds is difficult even under normal circumstances, and the fis tu la e constitute a case o f their own. Η. Cotton, D o cu m e n ta ry Letters o f R eco m m en da tion in L a tin fr o m the R o m a n E m p ire (Bei­ träge zur klassischen Philologie 132), Königstein, 1981, 28-33, with 28-9 for the date. 78 NERO’S ‘ARCHITECTS’ H ouse m entioned by T a citu s.19 This hypothesis and its im plications need to be briefly considered, even though, as w e now know , C eler w as not a rare nam e in the Roman world. Firstly, the fragmentary fam ily name o f C eler from Panopolis, f- ] iu s , is in agreem ent with the name w e have in the fis tu la . Then there is the question o f term in olo gy to co n ­ sider: is it co n ceiv a b le that the sam e man styles h im se lf a rc h ite c tu s in a letter from P anopolis and is called m a g iste r e t m a c h in a to r by Tacitus? T he outcom e o f m uch recent discussio n o f the tasks undertaken by Rom an a rc h ite c ti is that at tim es, an a rc h ite c tu s corresponded m ore c lo se ly to a m od em m aster-builder or construction en gin eer o f som e kind than to an architect.20 P recisely what a m a g iste r or a m a c h in a to r w as supposed to do is lik ew ise debated,21 but evid ently Tacitus thought such ‘jo b d escrip tion s’ w ould cover o verseein g the construction o f the D om us Aurea, in clu din g its sophisticated sp e­ cial spaces, as w ell as the diggin g o f a canal alo ng the Tyrrhenian shore. There seem s to be no reason w hy Severus or C eler m ight not have referred to th em selves as a rch itec ti. 4. T wo ‘arch itects’ w ith the sam e fam ily nam e The lead pipe from R om e, then, provides som e inform ation on the fam ily nam es o f N e r o ’s ‘architects’, hitherto unknown. Severus and Celer bore, in fact, the sam e n o m e n g e n tile , as show n by the expression d u u m [-J lio ru m in our inscription, C IL X V 7 3 9 3 . D ressel argued that on ly a very short nam e, m ost likely A eliu s or Iulius, w ou ld fit as a 19 20 21 Cotton (n. 18), 29, who notes that the hypothesis had previously been advanced by J. Rea; cf. the verdict of Eck (n. 3, 1997), 1049: ‘möglicherweise’. For discussions o f those professions in the Roman world that can be compared with the tasks o f modem architects, see G. Downey, ‘Byzantine Architects. Their Training and Meth­ ods’, B yza n tion 18, 1948, 99-118, esp. 108-12; P. Gros, ‘Statut social et role culturel des architectes (période hellénistique et augustéenne)’, A rch itectu re et so c iété de l 'a rchaïsm e g rec à la f i n de la rép u b liq u e ro m a in e (Coll. ÉFR 66), Rome, 1983, 425-52; W. Eck, ‘Magistrate, ‘Ingenieure’, Handwerker, Wasserleitungsbauer und ihr Sozialstatus in der römischen Welt’, M itteilu n g en des L eichtw eiss-In stitu ts f ü r W asserbau d e r T U B ra u n ­ sch w eig 103, 1989, 175-217; Μ. Donderer, D ie A rc h itek te n d e r sp ä ten röm isch en R ep u b lik u n d d e r K a iserzeit, Erlangen, 1996; W. Eck, ‘Auf der Suche nach Architekten in der römischen Welt’, JRA 10, 1997, 399-404 (reviewing Donderer); J.C. Anderson, R o m a n A r ­ ch itectu re a n d S o ciety, Baltimore, 1997; J. De Laine, ‘Organizing Roman Building and Space’, JR A 13, 2000, 486-92 (reviewing Anderson); Μ. Wilson Jones, P rin c ip les o f R o m a n A rch itecture, New Haven, 2000; ΤἩ . Howe, ‘Design Methods o f Roman Architects’, JRA 15, 2002, 465-8 (reviewing Wilson Jones). J. De Laine, ‘The Temple o f Hadrian at Cyzicus and Roman Attitudes to Exceptional Con­ struction’, P B S R 70, 2002, 205-30, esp. 216 holds that the function o f Severus and Celer was more than that o f ‘“master (-builder)” and “engineer”’; cf. Donderer (n. 20), 55 and in general n. 20 above. On the Greek expressions ἀ ρχιτέ κτων and μ ηχανικό ς see Α. Bernand, P a n d u désert, Leiden, 1977, 89-91, 96-8, 118-21, 191-2; Gros (n. 20), 428, 433 (including the related expression m achinator)·, and J.-P. Rey-Coquais, ‘Nom de metiers dans les inscriptions de la Syrie antique’, C C G G 13, 2002, 247-64 on αρχιτἐ κτων, τεχνί τη ς , and μ ηχανικό ς . Two m a ch in a to res are known from Rome and its surroundings, both freedmen: C. Baebius Musaeus co n lib ertu s in CIL VI 9533 = IL S 7727; L. Quinctius L. 1. Nicephorus in N SA 1953, 302 no. 70. CHRISTER BRUUN 79 com pletion o f the name ending in -liu s found on the fistula.22 That the p lu m b a r iu s , the manufacturer o f the lead pipe, is a Ὃ . Iuliu s’ m akes it sligh tly m ore probable that w e are dealin g w ith Iulii here,23 perhaps even G aii Iulii, and I w ill proceed on this assum ption, although it is not essentia l for the rest o f the argum ent.24 B efore con tinuing, it is worth m entioning a fact w hich has not been pointed out b e ­ fore: a certain C. Iulius Celer, or to be exact, C. lu i. C eler, appears in another f i s tu l a stamp, nam ely C IL X V 7 7 7 4 (= XI 3 6 8 5 c), found som ew hat north o f O stia, in La C hiaruccia, ancient Castrum N ovu m . The abbreviated g e n tilic iu m m akes it im p ossib le to determ ine whether a nom inative or a genitiv e is intended, that is, w hether the person w as the ow ner o f a v illa or a lo w ly p lu m b a r iu s . A s the tr ia n o m in a C. Iulius C eler can be found in other contexts as w e ll,25 it is risky to id entify the C eler at La Chiaruccia w ith the N eronian ‘architect’. A lead pipe stamp from V ercelli (P iem on te) presents a p l u m ­ b a riu s C. Iulius S everu s,26 but it seem s unlik ely that he is identical w ith N e r o ’s Severus. Since w e have now established that N e r o ’s m a g is tr i e t m a c h in a to r e s had the sam e g e n tilic iu m and perhaps w ere called Iulius Severus and Iulius C eler, it seem s lik ely that they com e from a com m on background. There are three a p r io r i p o ssib ilities as to h o w they cam e to bear the sam e fam ily name. W e m ight be d ealing w ith freeborn citizen s (in this case, presum ably brothers or som e other kind o f relations), or w ith fe llo w freedm en, or, as seem s less lik ely , w ith com panions w ho were g iv en Rom an citizenship at the sam e time and for the sam e reason, surely their professional sk ills.27 The idea that m ost o f the ‘architects’ and engineers in the R om an w orld cam e from the eastern parts o f the Em pire has enjoyed m uch support, but it has no foundation in the sources.28 It w ould be d ifficu lt to argue for an eastern origin o f Severus and Celer, based 22 23 24 25 26 27 28 Eck (n. 4), 263, suggested, exem p li g ra tia , Aelius, Iulius, Catelius or Catilius. An autopsy o f the stamp has unfortunately not been possible. The fistula is kept in the collection o f the Museo Nazionale Romano and is now in the process o f being prepared for scholarly access and investigation. In reality, there is surprisingly little correspondence between the family names o f private owners of water conduits and those o f the p lu m b a rii who manufactured them, see Bruun (n. 5), 348-50; C IL XV 7492a may be the only certain case. As we saw above, this discussion o f the g e n tiliciu m is not influenced by the identity o f the a rch itectus Celer from Panopolis, as only -ius can be read o f his family name. See, e.g., D. Μ. [.] Iu lio C eler[i] in A E 1988, 88 from Rome (= Via Im p eria le [Tituli 3], no. 209), and C IL VI 4579 (twice), 5662, 9405 iv.6, 840 = 30848. There is also the inscription C IL VI 14647 = 34085 C eleri N ero n is A u g u s ti Ι. According to various earlier reports, the text ended in the letters Α and Ο, which some have restored as a (rch itect)o , and Η. Brunn, G esch ichte d er g riec h isc h e n K ü n stler II2, Stuttgart, 1889, 232 suggested that the inscription refers to the person discussed here. If this Celer were our man, as a freedman o f Nero he would most likely be a ‘Τι. Claudius’ (unless freed by Gaius). Yet at C IL VI 14647 the edi­ tor comments that the stone shows no trace of, nor space for, the letters Α and Ο; similarly at C IL VI 34085. See G. Scalva, in L. Mercando (cd.), A rc h e o lo g ia in P iem on te II. L 'età ro m a n a , Torino, 1998, 94: C. lui. Sever. Vercel. fa c (it). In theory it is also possible that the identical family names are purely coincidental, and that the two had become Iulii independently o f each other, one being, e.g., a freeborn citizen, the other a freed slave. This was stated already by Brunn (n. 25), 225-26; recently underlined by Donderer (n. 20), 80 NERO’S ‘ARCHITECTS’ on what w e know o f them , as they both carry a Latin cogn om en , not a G reek o n e.29 If one looks at the know n instances o f a r c h ite c ti in the western part o f the em pire in order to find out what socia l status w ou ld be m ost likely for Severus and C eler, one can only say that there are enough both freeborn and freedm en am ong them to m ake these tw o alternatives equally p o ssib le.30 Unfortunately, the third person m entioned on the fistula, A ntonia L ivi (?), does not help us in determ ining the fam ily and socia l background o f Severus and Celer; she is otherw ise com p letely unknow n.31 5. From a w ater conduit on the Q uirinal to residen ce pattern s in R om e The find spot o f the f i s t u l a o f the ‘architects’ Severus and C eler is firm ly established by D ressel ( C IL X V a d lo c .) as ‘ad radices Quirinalis a l te a tr o N a z i o n a l e ’, i.e. halfw ay up the modern V ia IV N o v em b re.32 This location33* m eans that the lead pipe cannot have any connection to N e r o ’s G old en H ouse and that w e are dealing w ith a different project o f the tw o m a g is tr i e t m a c h in a to r e s . What, then, w as the purpose o f this conduit? W hy d oes the stamp also m ention a certain A ntonia Livi (?), and in what capacity did these three p eop le appear on the stamp? The m ost com m on interpretation is that the gen itiv e in a f i s t u l a stamp indicates ownership o f a co n cessio n o f public water, and that this private water grant w as m eant for a d o m u s ow ned by the proprietor o f the conduit. Thus in the recent L e x ic o n T o p o g r a p h ic u m w e find Antonia, C eler, and Severus duly listed as ow ners o f a d o m u s ? * N o t much effort has been d evoted to understanding what this m ultiple ow nersh ip o f the water conduit m ight, in fact, entail, and there has been no attem pt to qu estion the plausibility o f 29 -,0 34 73; similarly Eck (n. 20, 1997), 404; De Laine (n. 20), 488 (co n tra Anderson [n. 20], 14). The possible presence o f Celer at Panopolis at some point during his career does not, of course, prove his local origin, and it is, in fact, extremely unlikely that someone born in Egypt would have acquired a Latin cognomen there. Moreover, the Latin letter apparently written in his own hand — which is why we may presume that the language, too, must be his own — does not contradict the assumption that he came from a Latin speaking environment, even though he made a few orthographical errors and his style is not immaculate, as shown by Cotton (n. 18), 32-3. Donderer (n. 20), 71-6 (who does not discuss m a g istri and m a ch in a to res, 55, and therefore provides no basis for direct comparisons); Eck (n. 20, 1989), 188. See Raepsaet-Charlier (n. 4), 91 no. 75: ‘peut-etre épouse d ’un Livius, peut-etre sénatori­ ale?’ Strangely enough Dressel in C IL XV does not give any references to previous publications. It is unclear to me if someone else had reported the discovery in print before, nor do I know how Dressel learned of it. My frequent perusals of the pre-1901 volumes o f the B C A R and the NSA have not turned up any information on this particular lead pipe so far. The discov­ ery was evidently made in, or shortly before, 1888, when contruction work was undertaken on the site; see C. Bruun, ‘Velia, Quirinale, Pincio: note su proprietari di d o m u s e su p lu m ­ b a r i i A r c t o s 37, 2003, 27-48, esp. 32-3. The reference in the C1L was obviously to the late 19th-century Teatro Nazionale, not to the modern one, which is situated on the Viminal; see Bruun (n. 32), 33. There is room for confusion here, as can be seen in L T U R II (see the following note). See L T U R II, Rome, 1995, 33, 78, and 179. CHRISTER BRUUN 81 the com m only accepted explanation, once the im plications o f m ultiple ow nership have been w orked out. Α fair num ber o f f i s t u l a stam ps are know n in w hich m ore than one nam e appears.35 In about a quarter o f the over fifty such cases a w om an’s and a m an ’s nam e appear together (both in the genitive), and the com m on view is that we are dealing in these instances w ith a m arried couple.36 R egardless o f w hether this is consistently accurate (and I very m uch doubt it), this cannot be the situation here, for A ntonia is accom panied by tw o m en. T hen again, in about h a lf o f the cases w here we have m ultiple nam es on f i s t u l a e , it can be e s­ tablished that the persons nam ed are (close) relatives.37 F or Severus and C eler this m ight well be the case, but there is nothing to indicate that A ntonia (L ivi ?) belonged to the sam e family. N or do the three people seem to have shared one house, since A n to n ia’s name is apparently follow ed by a reference to her husband (although doubts as to w hether Dressel m anaged to decipher the dam aged portion o f the stam p correctly m ust rem ain).38 O ne possibility is that we are dealing w ith a jo in t lead pipe, installed w ith the purpose o f serving tw o properties (o r even three), perhaps m ade possible because m ore than one w ater grant was given by the em peror at the sam e tim e, to persons living near one a n ­ other.39 I f this w ere the case, A ntonia lived in proxim ity to Severus and C eler. C ontinuing this hypothesis, it is interesting that A ntonia holds a w ater grant in her ow n nam e, and cooperates with two other m en, w hile her husband does not appear as an ow ner o f the concession, only as the spouse o f A ntonia. I f the letters follow ing A n to n ia’s 35 'ft 37 38 39 Α fairly complete list was presented in Bruun (n. 5), 94 n. 75, to which a few other discover­ ies should be added. See C. Bruun, ‘Zwei Priscillae aus Ostia und der Stammbaum der Egrilii’, Z P E 102. 1994, 215-25, esp. 218 n. 14, and the following note. The theory was first presented by G. Barbiéri, O stia — fistole acquarie inedite o completate’, N SA 1953, 151-89, esp. 156-7 (= idem, S c ritti m ino ri, Roma, 1988, 290-1); accepted, e.g., by H.-G. Pflaum. Les ca rrières p ro c u ra to rie n n es éq u estres so u s le H au t-E m p ire rom a in III, Paris, 1961, 996-7, and C.R. Whittaker in his Loeb Classical Library edition of Herodian, vol. I, London, 1969, lxxvii, 354 (regarding the marriage o f the equestrian L. Didius Marinus to Cornificia, daughter of the emperor Marcus), 354 (the marriage of Vibia Sabina to L. Aurelius Agaclytus); also in G. Alfoldy, ‘Die Stellung der Ritter in der Führungsschicht des Imperium Romanums’, C h iro n 11, 1981, 169-215, esp. 195-6. Both marriages were also recorded by Raepsaet-Charlier (n. 4), 261 no. 294 and 622-3 no. 800; and by D. Kienast, R öm ische K a iserta b elle 2, Darmstadt, 1996, 140, albeit with a question mark. Some ten fis tu la stamps containing one male and one female name were listed in Bruun (n. 5), 94 n. 75, to which can now be added M N IR 58, 1999, 33 (see at n. 56 below) and M D A IR 108, 2001,325. See Bruun (n. 5), 94 n. 75 for a list of 23 such stamps. There are a handful o f stamps in which a woman appears as wife o f her husband: C IL XV 7421 C aepiae P rocu la e Μ R eg u lfi uxoris)·, XV 7440 C o rneliae Ta uri f. Τ. A xi; XV 7441 C orneliae L. f. Volusi S a tu rn in i p .; XV 7549 S ulp icia e Q. f . P ra etexta ta e C ra ssi; XV 7580 [-] M . f P riscae C. R u fi[n i? uxoris]; B C A R 1941, 191 no. 32 [ P a ju lla e F austi. The Roman jurists mention a situation where two parties share the same conduit (eodem rivo ) — see Dig. 43.20.5.1 (Iulianus, libro q u a rto ex M inicio) — although it probably con­ cerned irrigation. According to Frontinus (a q . 109.4-5), water grants to multiple beneficiaries were possible, but he also notes that matters relating to water grants to so cii were sometimes unclear. 82 NERO’S ‘ARCHITECTS’ name denote her husband, she can be added to the six senatorial w om en w h o nam e their husband in fi s tu l a stam ps, although these other w om en (m ostly dated to the first century C E )4041alw ays appear w ithout any further partners. That Severus and C eler should have been awarded the ius a q u a e d u c e n d a e in the city o f R om e is not surprising. Even i f the f is tu la is to be dated to a period after N e r o ’s reign, sk illed construction experts w ould continue to be hig hly regarded. W e know that one o f the em peror C la udiu s’ physicians w as giv en a water grant, and this in dicates that experts c lo se ly con nected w ith the court were giv en specia l privileges.'" The collocation o f the nam es Severus and C eler in the stamp points to a shared w ater co n cession , w hich m eans that they shared the sam e house, or at least had neighbourin g properties. A ccordin g to Frontinus, a water grant w as alw ays given for a sp ecific prop­ erty, and it w as not perm itted to conduct the water elsew here (aq. 10 9 .6 ). Α considerable number o f prom inent senators also appear in stam ps o f this type, poin tin g to a different and not infrequent pattern o f ownership, as show n by stam ps such as I I (d u o ru m ) Q u in tilio ru m C o n d ia n i e t M a x im i ( C IL X V 7 5 18) or S e x tio ru m T o rq u a ti e t L a te r a n i ( C IL X V 7 5 3 6 ).42 It m ay be worth con sid erin g whether this also indicates a sim ilar pattern o f resi­ dence. The fact that w e have over tw enty fis tu la e from R om e with such ‘fam ily stam ps’, i.e. over 10% o f all the persons m entioned in f is tu la stam ps appear in su ch a con text — and som e sim ilar stam ps have been found elsew here in Italy43 — has not b een taken into account by m od em scholarsh ip on residence patterns in the capital and the organization o f the Roman h ou se.44 I f all these stam ps with m ultiple nam es indicate resident ow ners 40 41 42 43 44 See n. 38 above for the names. The datings are provided by Eck (n. 4), 259-77 {p a ssim ). Eck (n. 4), 249 (C. Stertinius Xenophon, physician o f Claudius, in C IL XV 7544). Other similar stamps are, e.g., C IL XV 7438, 7472, 7511; B C A R 1905, 294. For other refer­ ences, see Bruun (n. 5), 94 n. 75. The are twenty-one stamps from Rome in which people who are clearly or most likely rela­ tives appear: C IL XV 7393 (our stamp), 7425, 7438, 7472, 7481, 7509 (?), 7510, 7511, 7515, 7517, 7518, 7523, 7525, 7536, 7545; B C A R 1905, 294; B C A R 1906, 35 no. 3; B C A R 1907, 349; B C A R 1915, 220; C a rta Arch. R om a II, 68; E p ig ra p h ica 1951, 23 no. 34. Else­ where in Italy we find seven more: C IL Χ 1905; A u so n ia 6, 1911, 48; Eph. Ep. VIII 376; N SA 1953, 165 no. 22; RP A A 64, 1991-92, 247; Suppl. It. 2, 52 no. 16; S upp l. It. 3, 34 no. 21. There is no analysis along these lines in connection with the several fis tu la e that name senatorial Neratii, in L T U R II, 144; nor is there any in recent scholarship on aristocratic living in Rome, such as F. Guidobaldi, ‘L ’edilizia unifamiliare nella Roma tardoantica’, in Α. Giardina (ed.), S o cietà ro m a n a e im pero ta rd o a n tico II, Roma, 1986, 165-237, 446-60; W. Eck, ‘C um d ig n ita te otium . Senatorial d o m us in Imperial Rome’, S C I 16, 1997, 162-90 (although on 164-5 there is, exceptionally, a brief reference to occasional multifamilial residences); L. Chioffi, ‘Sulle case delle élites a Roma e dintomi. Supplemento al L exicon T op og ra ph icum U rbis R o m a e ', B C A R 100, 1999, 37-52; J.-P. Guilhembet, ‘Les résidences aristocratiques’, P a lla s 55, 2001, 215-41. See too Α. Wallace-Hadrill, ‘The Streets o f Rome as the Representation of Imperial Power’, in L. De Blois e t al. (eds), The R ep resen ta tio n a n d P ercep tion o f R o m a n Im p eria l P o w e r (Impact o f Empire 3), Amsterdam, 2003, 189206, who on 192 refers to medieval Rome and Genoa: ‘The characteristic o f such a city is the to rre as the heart o f the power of the family clan, surrounded by a local maze o f narrow streets, the co n tra d a or patch the family controls’. On this phenomenon in medieval Rome, see H. Broise and J.-C. Maire Vigueur, ‘Strutture famigliari, spazio domestico e architettura CHRISTER BRUUN 83 o f houses, w e m ay, in fact, be dealin g with a situation rem iniscent o f m edieval R om e, where certain neighbourhoods were populated by m any branches and generations o f a powerful fam ily.'15 In fact, w e happen to kn ow that in the first century B C E M arcus C icero and h is brother Q uintus lived in neighbouring d o m u s at one point in tim e (C ic. A tt. 4 .3 .2 ), but since the tw o m en were h o m in e s n o v i from Arpinum, it is clear that the area w as not tra­ ditionally inhabited by the T ullii, nor did the tum ultuous events during the late R epublic allow the fam ily to establish a long term base on the slo p es o f the Palatine.'16 It has been pointed out that real estate changed hands very frequently am ong aristocratic fam ilies, w hile the com positio n o f the upper cla ss steadily underwent changes through social m o ­ bility, so that, in practice, permanent concentrations o f one fam ily in on e quarter w ere uncom m on.454647 Indeed our sources, w hich, to be sure, are very incom plete, g iv e very fe w indications o f such fam ily neighbourhoods. Perhaps tw o generations o f the N um m ii lived on the sam e site in the fourth century48 and the F lavii inhabited a certain quarter o f the Quirinal before V espasia n becam e em peror.49 There are also several fis tu l a e and other sources nam ing Neratii from p la ces not too far apart (although stretching over several city blo cks), and it has b een suggested that a few other fam ilies m ay have inhabited the sam e palace during several generations.50 The ‘D om us dei V a lerii’ on the C aeliu s sh o w s inhabitation during four centuries (I-IV CE), but the written sources on ly refer to certain senators o f the fourth century.51 O f the know n inscriptions relating to fam ily ow nership o f houses, the one m entionin g the d o m u s A r ip p o ru m e t U lp io ru m V itio r u m com es clo sest to our f i s tu l a stam ps, but 45 46 47 48 49 50 51 civile a Roma alia fine del Medioevo’, in Ρ. Fossati (ed.), S to ria d e l l 'arte ita lia n a 12, Torino, 1983, 97-160, esp. 114-41. J. Hillner, 'D o m u s, Family and Inheritance: the Senatorial Family House in Late Antique Rome’, J R S 93, 2003, 129-45, e.g. 130, argues outright against the existence of enduring topographical concentrations. Α good example is the Cenci family, which established its seat in the Campus Martius, still identifiable today next to the piazza with the same name near the synagogue; see Μ. Bevilacqua, Π M onte d ei Cenci. Una fa m ig lia rom a n a e il su o in sed ia m en to urba n o tra m ed io evo e d età b arocca, Roma, 1988. The Monte Giordano in the Campus Martius across from Castel S. Angelo is another example. The main building is now known as Palazzo Tavema but was once a centre for the Orsini family; see S. Carocci, ‘Baroni in città. Considerazioni sull’insediamento e i diritti urbani della grande nobiltà’, in É. Hubert (ed.), R om e a u x X IIIe et X IV e siècles (Coll. ÉFR 170), Rome, 1993, 137-73, esp. 142, 153. See L TU R II, 202-04 (Ε. Papi) on their houses. The younger Marcus Cicero survived, but his father’s house ended up in other hands. The Bruttii Praesentes are cited as an exception by Eck (n. 44), 188-90. L T U R II, 146-47 (F. Guidobaldi); Eck (n. 44), 190. L T U R II, 102-04 (W. Eck); Eck (n. 44), 165. As noted by Eck (n. 44), 188, who mentions the Acilii Glabriones and Scipiones Orfiti in addition to the Neratii. The issue here, however, seems to be the longevity o f these g en te s in the Senate rather than the existence o f permanent family headquarters in Rome. As the L T U R II, s.v. ‘domus’ shows, nothing in particular is known about the residence of the Scipiones Orfiti during the empire, while for the Acilii the information is restricted to the consul o f 438 CE (99-100). Guidobaldi (n. 44), 186-88; L T U R II, 207 (F. Guidobaldi). 84 NERO’S ‘ARCHITECTS’ these tw o fam ilies are otherw ise unknow n and seem to be o f non-aristocratic status.52 O ther sources, ju rid ic a l and literary, give very little support to the idea th at adult sons w ould live with their fathers, or, to cite G am sey and Sailer, ‘th at adult brothers w ould share a com m on household as a c o n s o r tiu m ' , except in cases w here the fam ilies w ere p oor.53 C onsequently, it is perhaps understandable that our fis tu la e , w hich list m ultiple fam ily m em bers, have not prom pted a revision o f the current view s on aristocratic living in Rome. Y et, surely it is desirable that the f i s tu la stam ps nam ing brothers and close relatives also be considered in this context. W e have to ask w hat it m eans w hen the senatorial C ornelii (Fronto and Q uadratus), the Q uintilii, the Petronii (S ura and M am er­ tinus), the N eratii, and the Sextii appear in such stam ps. In addition, individuals o f low er status, such as the Iulii C efalü, the Iunii, and others, also need to be taken into account when pondering the social and civic significance o f these stam ps. The next and final section w ill explore w hether approaching these stam ps from a dif­ ferent angle, i.e. assum ing that these fis tu la e do n o t refer to the residence o f the individuals m entioned, m akes better sense. 6. A ltern atives: no residen ce, no domusl It is w orth exploring w hether in the case o f A ntonia and the tw o ‘a rch itects’ a different explanation o f the purpose o f the conduit w ould be m ore persuasive. In recent years, I have argued that a nam e in the genitive on a f is tu la need not alw ays refer to the ow ner o f a d o m u s ,54 and som e new discoveries have provided further p ro o f for this claim . For instance, the stam p R o i H ila rio n is (C IL X V 7522) denotes a p lu m b a r iu s , not the ow ner o f som e p roperty,55 w hile the stam p Μ C o rn e li S e c u n (d i) et S e r g ia e P a u la e , found in the courtyard o f the C asa di D iana at O stia, obviously refers to the ow ners o f an ap art­ m ent building, not a d o m u s .56 O ther conduits with a nam e in the genitive m ay have supplied w ater to industrial establishm ents, in particular baths.57 T here are still other possibilities, although they are not likely to have m uch bearing on the present case.58 52 53 54 55 56 57 58 See L T U R II, 37 (F. Guidobaldi) for the text and the dom us. Ρ. Garnsey and R. Sailer, The R o m a n Em pire. E conom y, S o c ie ty a n d C u lture, London, 1987, 129; similarly R. Sailer, P atriarchy, p ro p e rty a n d d ea th in the R o m a n fa m ily , Cam­ bridge, 1994, 82: ‘The jurists ... did not envisage the possibility o f a joint household, either one headed by two brothers in co nsortium or one of three generations including daughtersin-law’. On a different possible pattern in aristocratic families labouring under economic hardship, see J. Crook, 'P a tr ia p o te s ta s ', C Q 17, 1967, 113-22, esp. 117. Bruun (n. 5), 81-95; Bruun (n. 35); C. Bruun, ‘Private Munificence in Italy and the Evi­ dence from Lead Pipe Stamps’, in H. Solin et al. (eds), A cta C o llo q u ii E p ig ra p h ici L atin i (H elsingiae 1991), Helsinki, 1995, 41-58; idem, JRA 10, 1997, 389-98, esp. 396-8; idem, JRA 13,2000, 498-506. Bruun (n. 32), 36-42. Α. Marinucci, ‘La distribuzione dell’acqua nella cosiddetta Casa di Diana’, M N IR 58. 1999. 32-35, esp. 33. Bruun (n. 35); Bruun (n. 54, 1995), 48-52; R. Geremia Nucci, ‘Le Terme del Faro di Ostia. Nuovi dati provenienti dallo studio delle fis tu la e ', A rch C la ss 51, 1999-2000, 383-409. The genitive sometimes tells us who sponsored a particular public building, but this is not very likely in the centre of Rome, where the emperors jealously monopolized public build­ ing. CHRISTER BRUUN 85 O ne aspect that has been neglected is the size o f our conduit. D ressei ( C IL X V 7 3 9 3 , a d lo c .) gives no exact m easure, but he w rites ‘m a g n i m o d u li', w hich, in his language, m eans a diam eter o f at least 15 cm. As far as R om an w ater pipes go, this is a very large w ater pipe indeed. N orm ally, pipes intended for private individuals have diam eters as small as 5 cm or even less, and a diam eter o f 15 cm does not indicate th at a f i s tu l a is the com bination o f three com m on concessions o f 5 cm w ater pipes. T he area (Α ) o f the pipe is decisive for the am ount o f w ater a pipe can hold, and using the form ula A = n r 2 we get an area o f 176 cm 2 for the pipe o f A ntonia and com panions, w hile a 5 cm pipe has an area o f ca. 20 cm 2. T he larger pipe delivers over eight tim es as m uch w ater as the sm aller one.59 Today it is clear that all the largest fis tu la e found in R om e carry stam ps o f the e m ­ peror and w ere either w ater m ains or supplied im perial o r public structures. O nly very few so-called private pipes have a diam eter sim ilar to the m a g n u s m o d u lu s o f our f i s ­ tu la .60 I also suspect that in at least som e o f these cases we are dealing w ith pipes that in fact supplied public structures o f som e kind.61 T he question now is w hether the urban structure at the foot o f the Q uirinal could accom odate tw o (o r even three) d o m u s w ith extensive gardens — how else could such enorm ous am ounts o f w ater be used in a p ri­ vate context? — o r w hether we should surm ise that the w ater had another purpose. A jo in tly ow ned bath is one possibility, and there w ould be nothing strange in a m arried w om an investing on her own (at any rate if we assum e that she w as s u i iu ris ) in a bath, for instance. T hen again, the fact that Severus and C eler w ere fam ous builders, also in­ terested in w aterw orks — probably not ju s t canals (Tac. A n n . 1 5.42), for one w ould im agine that the delights w hich the Dom us A urea w as to offer also included hydraulic m iracles - raises further questions. W ere it not for the nam e o f A ntonia, w hich seem s odd in such a context (although she m ay not have been a senatorial lady at all, but an entrepreneur), one m ight suggest that their nam es are an indication that they w ere in­ volved in som e project m eant to im prove the urban infrastructure. N o doubt there w as a great need for building in this region after 6 4 CE, and not ju s t under N ero. T he great fire had caused such destruction that there m ust have been constant construction w ork in m any parts o f R om e throughout the Flavian era. * * * One can only hope that som e unexpected archival find w ill one day shed m ore light on the circum stances o f the discovery o f the f is tu la o f Severus, C eler and A ntonia. W hat can be established w ith certainty is that in C IL X V 7393 we find a hitherto undiscovered reference to N e ro ’s m a g is tr i et m a c h in a to res . T he interpretation one should give to their jo in t appearance, in the com pany o f one A ntonia, is m ore difficult to pin dow n. T he p o s­ sibility that we are dealing with the com m on residence o f all three m ay be safely excluded, and I find it im probable that the pipe w ould have supplied tw o or three sep a­ rate buildings inhabited by these individuals. It is m ore likely that w e are dealing with a 59 60 61 The volume also depends on the water pressure and the interior surface o f the conduit, see Α.Τ. Hodge, R o m a n A q u ed u cts a n d W ater S up p ly, London, 1991, 215-44. Bruun (n. 5), 138. See C. Bruun, ‘Grapti Aug. 1. praedium’, in Α. La Regina (eel.), Lexico n T op og ra ph icu m U rbis Rom ae. S u b u rb iu m III, Roma, 2005, 39-40. 86 NERO’S ‘ARCHITECTS’ join tly ow n ed p iece o f real estate, the result o f an investm ent or perhaps an inheritance. I f this last suggestion seem s less attractive, w e can alw ays return to the hypothesis o f ‘clan resid en ces’ in R om e, even though previous scholarship has found very little trace o f such an urban phenom enon in our sources for im perial R om e. U niv ersity o f Toronto